content-locale-humanize-th · diff
git:20260918.40e721d to git:20261013.a69e0bb
20 added, 7 removed. Audit A to A.
---
name: content-locale-humanize-th
title: "Thai Native-Fluency Calibration"
description: "Per-language calibration anchors for detecting AI-slop and translationese in Thai (th) target text; auto-loads alongside content-locale-humanize when the target language is Thai."
license: Apache-2.0
- compatibility: "Extends content-locale-humanize's AI-Slop & Translationese dimension. Load both together — this file has no rubric of its own."
+ compatibility: "Calibration data only; no tools required. Network access helps for usage checks."
domains: content
rules:
- content(thai)
- match(\bth-TH\b)
- match(\bThai\b)
- match(ภาษาไทย)
- semantic(check if this Thai translation sounds native)
- semantic(eliminate translationese from this Thai text)
---
## Overview
- Sourced calibration anchors for Thai, feeding `content-locale-humanize`'s AI-Slop & Translationese dimension. This is a calibration aid, not the checklist — reason natively beyond it (see the core skill's "why structure, not word lists" section).
+ Calibration anchors and current norms for Thai (th-TH), feeding content-locale-humanize's dimensions. A calibration aid, not the checklist; reason natively beyond it.
- ## Instructions
+ ## Rules
- Sourcing confidence: weakest of any language in this set. A targeted search turned up Thai-specific AI-detection tooling but no independently-documented word-level or phrase-level AI tells — don't invent a plausible-sounding list to fill the gap.
+ ### AI tells
- Lean entirely on the structural checks in `content-locale-humanize` for Thai: source-syntax interference (Thai has no articles, no verb tense inflection, and a different clause-topicalization pattern than English — literal English sentence structure carried into Thai is unusually visible), sentence-rhythm clustering, connector-pileup, and paragraph-opener templating. Thai's lack of spaces between words and different clause-boundary conventions also make burstiness/rhythm judgments harder to make by eye — read for natural clause-break rhythm rather than word-count.
+ Sourcing confidence: the weakest in this set. Searches found Thai AI-detection tools but no independently documented word- or phrase-level AI tells. Don't invent a plausible list to fill the gap; if a sourced list appears, add it here.
- If a genuine Thai-specific AI-tell word list becomes available, add it here — this file exists as the attachment point.
+ ### Translationese from English
+ Lean on the structural checks, which are unusually visible in Thai:
+
+ - English sentence structure carried over: Thai has no articles or tense inflection and topicalizes differently.
+ - ถูก-passives copied from English; native Thai reserves ถูก mostly for adverse events and otherwise prefers active or topic-first sentences.
+ - Pronouns Thai would drop (คุณ, เขา, มัน) and stacked การ/ความ nominalizations mirroring English abstract nouns.
+ - Rhythm is hard to judge by eye without word spaces; read for natural clause breaks rather than word counts.
+
+ ### Current norms
+
+ - No spaces between words; a space marks a clause or sentence break, full stops are not used in running text, and English-style comma density reads as translated.
+ - Polite particles (ครับ/ค่ะ) are gendered: in UI and documentation either avoid them or follow the brand voice, consistently.
+ - Spelling and transliteration follow the Royal Society of Thailand; formal and government contexts use Buddhist Era years (CE + 543).
+
## References
- No solid per-language sourcing found as of this writing; see Overview.
+ Reviewed 2026-10-12. No solid per-language AI-tell sourcing found. Norms: Royal Society of Thailand (ราชบัณฑิตยสภา) dictionary and transcription rules.